Scottish Government

CHAPTER FOUR: INTERVENTIONS FOR RURAL ROADS

Education, training and publicity campaigns

Background

4.1 The main aims of driver education, training and publicity campaigns are firstly, to teach novices how to drive and secondly, to improve standards of existing drivers. Their use has two important benefits compared with engineering measures:

• it addresses the root cause of the problem - the interaction between road user and road layout or traffic situation;
• any resulting safety improvements are not likely to be offset by reductions in capacity, which can occur as a result of geometric changes to improve safety

4.2 As a consequence, the potential benefits of effective driver training, education and publicity are very great. However, the direct benefits can be difficult to measure. The drink-driving campaign is an example that has substantially reduced the incidence of drink-driving over the years. The Scottish 'Foolsspeed' campaign is recognised as having had a positive effect on attitudes to speeding (Stead et al, 2002) although it is also recognised that interventions designed to change people's attitudes have had little observed impact on driving speeds (e.g. Parker et al, 1996). Stead et al (2002) note that advertising alone is unlikely to impact on behaviour unless part of a multi-faceted strategy including enforcement and environmental changes.

4.3 Very little information exists on education, training and publicity campaigns which are specific to rural areas. However, several initiatives (especially publicity campaigns and driver training) have targeted issues and groups which are relevant to rural areas and these are discussed in the following sections.

Education

4.4 Education in road safety begins in schools and may be aimed at various user groups including, for example, child pedestrians, cyclists or young drivers (e.g. Driving Standards Agency education package). Many of these initiatives are facilitated or delivered via local authority or police road safety units.

4.5 Christie et al (2002) carried out a survey of local authority policy with respect to road safety applicable to children in rural areas. They found a number of common policies and strategies although none were targeted at or exclusive to children in rural areas. The education initiatives highlighted include:

• Classroom-based road safety education in the National Curriculum
• Theatre in Education campaign against speeding involving newly qualified drivers
• Children's Traffic club for 3 to 4 year-olds
• Young driver education courses
• 'Crucial Crew' - incorporation of road safety education into broader events on safety targeted at 10-11 year olds
• Cycle training schemes

4.6 Christie et al (2002) point out that education initiatives are rarely evaluated and even then, the evaluation is usually informal.

Publicity campaigns

4.7 Some campaigns are aimed at the general public and focus on issues such as speed or fatigue whilst others target particular categories of driver, for example young males, who are disproportionately involved in accidents. Like education initiatives, the effects of campaigns appear to be rarely evaluated.

4.8 The THINK! Campaign recently focussed on the dangers of driving at inappropriate speeds on rural roads. It emphasised the potential problems encountered by drivers travelling too fast on these roads and highlighted that rural roads can present unforeseen hazards such as blind bends or animals in the road. The campaign included a radio advert aimed at younger drivers which highlighted the dangers of driving too fast on rural roads (DfT, 2004c).

4.9 Quiros and Shaver (2003) report on initiatives to reduce crash fatalities on rural roads in the USA. They point out that most funding is given to projects in higher population areas but that there are some initiatives specifically targeted at rural roads. For example, they point to 'Partners for Rural Traffic Safety Action Kit Guide' which was a publication designed to educate rural community leaders on how to promote a seat-belt use campaign and measure its success. It should be noted that seat belt laws in the USA vary from state to state and seat belt usage is typically much lower than in Scotland, at around 75% (Glassbrenner, 2003).

4.10 In a survey of education literature regarding driver sleepiness mainly in the UK, Australia and the USA, Flatley and Reyner (2000) found three main approaches to driver education. These were:

• Education of the general public about the dangers of driving whilst sleepy
• Targeting high risk groups using appropriate material and;
• Education of other opinion formers (such as employers, teachers, law enforcement personnel)

4.11 They also identified some innovative schemes in the USA and Australia including roadside restaurants encouraging drivers to stop with free coffee and the novel audio cassette and booklet entitled 'Wake Up!' which was distributed to radio stations and drivers in Australia.

4.12 Recently in the UK, the Department for Transport's 'THINK! Campaign has included radio advertising to alert drivers to the dangers of driving whilst tired.

4.13 The issue of older drivers turning right has been highlighted as a potential problem on some rural roads. For example, Faulkner (1975) found that the risk of an accident at the junction of a minor road with a rural major road increased with the age of the driver. Hughes and Amis (1996) also found a 'significant association between car driver age and manoeuvre which was independent of gender'. In particular, they found that right turn manoeuvres were riskier for drivers aged over 60. However, unpublished research for the Highways Agency found that it was more difficult to target older drivers through publicity campaigns than younger ones.

4.14 Australia is one country where rural road safety issues have been recognised with a national rural road safety strategy (Federal Office of Road Safety, 1996). This plan has included several public education programmes specifically targeted at issues affecting rural areas. These have included fatigue, drink driving and seat belt usage (all major issues in rural road accidents in Australia). The latest campaign related to speed and targeted young males aged 17 to 25. The campaign was based on research which suggests that young males are less likely to respond to messages which warn of death or injury (either to themselves or to others) but more likely to respond to messages that they might lose their licence and the likely consequences of this ( ATSB, 1999).

Driver training/ licensing

4.15 There is very little literature which deals with driver training or licensing specific to rural areas. However, there is a great deal of literature relating to these topics in general and some of the more relevant papers are discussed here.

4.16 The driving test is regarded as the key element in driver training. It continues to be improved over time and has included a written, as well as a practical component for a number of years. Recently, video clips to improve learner drivers' hazard awareness (Grayson and Sexton, 2002) have been adopted as an integral part of the test.

4.17 Unpublished research by TRL for the Highways Agency directed at young male drivers was intended to address their lack of driving experience and the likelihood that they would be over-confident in their own abilities. A video-based hazard perception training package developed by the Driving Standards Agency as an extension of those now included in the driving test was used with a group of newly qualified drivers. The clips were designed to stimulate discussion within the group and to demonstrate good practice. A control group received no training. However, because accidents are rare events and can occur at any point in the road network, it was not possible to conclude how effective the training was.

4.18 A number of organisations offer post-test on-road training in the UK:

• IAM (Institute of Advanced Motorists)
• RoSPA (Royal Society for the Prevention of Accidents)
• Drive & Survive UKPLC
• Driving Services
• DriveTech
• BSM (British School of Motoring)

4.19 However, there is no legal requirement for drivers to undertake post-test training and, therefore, it is undertaken by only a small proportion of drivers. The novice driver scheme called 'Pass Plus' is linked to reduced insurance premiums which may increase take-up.

4.20 Traditionally, mastery of traffic situations has been the main component of driver training (e.g. Hatakka et al, 2002). However, it is the case that excellent skills for this are not necessarily enough for safe driving (see Lonero et al, 1995). By concentrating on the technical aspects of driving and increasing the self-confidence of novice drivers, the more skilled drivers may simply drive faster, overtake in heavier traffic or listen to the radio (Evans, 1991). There is evidence to suggest that it is attitude rather than skill that is related to crash involvement, assuming that driving skill has reached a minimum standard ( DETR, 2000 and see Lonero et al, 1995). Several personality characteristics have been associated with unsafe driving, the most prominent of these being hostility/aggression, thrill-seeking and impulsiveness. To avoid this, driver training should address driving style and include awareness of personal skills and their limits.

4.21 Various European initiatives have been undertaken by CIECA (the international commission for driver testing authorities) with regard to driver training ( www.cieca-drivinglicense.org ):

• BASIC
• Advanced
• GADGET
• NovEV

4.22 Advanced is an EU project that undertook a study of post-licence driver and rider training. It describes and analyses voluntary, post-licence training and makes a series of recommendations on how to improve such training. It emphasises the importance of avoiding overconfidence amongst trainees and indicates how training can be more effective and balanced. The CIECA website contains a report on the project with guidelines on how to evaluate the effects of training on participants and a Risk Awareness Database of exercises which can be used in the training. The report highlights the lack of relevant research into the effects of post-licence road safety training.

4.23 Advanced is accompanied by a sister project called BASIC on new training methods for (pre-licence) learner drivers.

4.24 The GADGET project developed a matrix depicting four levels of driver behaviour and the focus needed on each of these levels to make "a good driver". The matrix is based on the assumption that the driving task may be described as a hierarchy such that abilities and preconditions in a higher level influence the demand and preconditions on a lower level. The levels are:

• Goals for life and skills for living - lifestyles, social background, gender, age and other individual preconditions influence attitudes, driving behaviour and accident involvement
• Goals and context of driving - why, where, when and with whom driving is carried out e.g. time of day, whether fatigued, trip purpose
• Mastering traffic situations - adjust driving in accordance with traffic e.g. at junctions
• Vehicle manoeuvring - control of vehicle, injury prevention systems e.g. seat belts, air bags

4.25 NovEV is a successor to the Advanced project which included recommendations for countries wishing to introduce compulsory "2nd phase" training for novice drivers (after passing the driving test), including advice on the methods used to evaluate the effects of the training on participants.

4.26 Williamson (2003) reviews the specific issues associated with young drivers and in particular covers Graduated Licensing Schemes ( GLS). She suggests that the issue of inexperience is well tackled by GLS but other issues such as risk-taking may not be specific to young drivers. She recommends that more work is carried out on finding the best ways of tackling so-called 'problem drivers' who may or may not be young drivers.

Fleet driver training

4.27 As highlighted in Chapter 3, driver fatigue is a recognised factor in rural accidents. Large goods vehicle drivers commonly experience driving pressure from tight schedules and long hours, and these may increase at particular times of year (Jackson, 2004). There are often no official company rules or guidance related to driving safety. A prevalent attitude in the small haulage companies is that there is no need for rules or training and no point in accident reporting/feedback (DfT, 2004d). In addition to basic training, a Safety Management System (see e.g. British Standard 8800 (1996) 'Guide to Occupational Health and Safety Management Systems') is desirable to manage road safety for fleet drivers.

Engineering measures

4.28 In contrast to the education, publicity and training sectors, there is a very large volume of literature on engineering interventions specifically for rural roads. Many of the measures and evaluations are described in the IHT guidelines ( IHT, 1999) and summarised in the Road Safety Good Practice Guide ( DTLR, 2001). Advice note TA85/01 ( DMRB 6.1.3) provides useful guidance on the assessment and design of minor improvement schemes for trunk roads.

4.29 The OECD (1999) highlights that safe road design must take into account the capabilities and limitations of road users. This safe design philosophy has two main strands: making human error less likely and forgiving human error if it occurs.

Rural speed management

4.30 As has been highlighted earlier in Chapter Three, speed management is a key issue in rural road safety. The setting of speed limits, appropriate to the needs and functions of the specific road section, is fundamental to the speed management philosophy.

4.31 The Department for Transport investigated the potential for setting a new rural road hierarchy (as suggested by a number of studies) but concluded that a new hierarchy would be costly (both financially and environmentally) and the benefits would take a long time to realise. However, this decision did not preclude the setting of more consistent speed limits within existing infrastructure constraints (Lynam et al, 2004).

4.32 Current guidance on the setting of speed limits has been interpreted in different ways by local highway authorities with the result that, over the last few years, different decisions have been made on the choice of appropriate speed limits for rural roads in different parts of the country (Lynam et al, 2004).

4.33 Lynam et al (2004) set out a new method for assessing and deriving speed limits on rural single carriageways. This divides roads into two groups: an upper tier (which would include most A and B roads) and a lower tier (which would include most minor roads). For the upper tier roads, a speed limit of 50mph should initially be considered. For 'higher quality' roads, a speed of 60mph may be appropriate provided that the accident rate is below a set threshold. For lower tier roads, a initial speed limit of 40mph should be considered with the potential for a 50mph limit on 'higher quality' roads, again provided that the accident rate falls below a certain threshold. Where roadside development dictates, 30mph and 40mph limits should be provided. Lynam et al (2004) predict accident savings of the order of 25% on upper tier roads and half this on lower tier roads.

4.34 Based on the findings of Lynam et al (2004), the Department for Transport and the Scottish Executive are currently consulting on a new process for the setting local speed limits including, in particular, on rural single carriageways (DfT, 2004e and Scottish Executive, 2004d). If this guidance is adopted and implemented, it should result in more consistent and understandable speed limits throughout the country.

4.35 However, the new guidance makes it clear that

"speed limits should be considered as only one part of rural safety management. The first priority where accident rates are high should be to seek cost effective improvements to reduce these rates, targeting the particular types of accidents taking place."

It points to research ( TRL Limited, 2004) which gives guidance on how to assess rural road safety and identify types of route or intervention measures which may be appropriate to reduce speeds and accidents along the route. The TRL guidance suggests separate intervention levels for single and dual carriageways (see also Barker et al, 1999) and points to the use of engineering measures which can be implemented to manage speed in rural situations (see DTLR, 2001). Some of the more important or innovative engineering interventions applicable to the rural environment are discussed in the following sections.

Vehicle-activated signs

4.36 Vehicle-activated signs light up only for drivers exceeding a pre-set speed. They may display a speed limit or advance warning of a hazard (e.g. a bend or a crossroads junction). Reductions in mean speeds of 3-6mph have been observed following the introduction of such signs on the approaches to bends, junctions or a speed-limit change (Barker, 1997, Farmer et al, 1998, Webster, 1995, Winnett et al, 1999, Winnett and Wheeler, 2003), depending on the traffic flows and before speeds. Winnett and Wheeler (2003) found that the signs led to a substantial accident reduction.

Countdown signs

4.37 Barker (1997) found that countdown signs did not affect mean on-road speeds. However, Pyne et al. (1995A and B) using a driving simulator, found they were more effective than a speed limit sign alone. Wheeler & Taylor (1999) recorded large reductions in mean speed when countdown signs were used in conjunction with other measures having a high visual impact at a village gateway. Similar results were reported by Taylor et al (2002b) using a driving simulator.

Road markings and surfacing measures

4.38 White edge lining is recommended to delineate the edge of the road and is used in conjunction with a hardstrip on major roads. Its use may lead to higher speeds at night because it is easier for drivers to see the line of the road ahead. Hatching can also be used at the edges of the road to reduce lane width. Longitudinal red strips with hatching on the edges and centre of a rural single-carriageway road were found to be effective in reducing mean speeds on a driving simulator by up to 5.6 mph (Taylor et al, 2002b).

4.39 Raised rib edge markings, which alert the driver if the vehicle crosses the line, have been used on motorways for a number of years and may be beneficial in reducing sleep-related accidents.

4.40 Centre lines (and lane lines) delineate lanes and again aid the driver to see his path ahead. The former are also used for hazard marking (e.g. use of double white lines at bends).

4.41 The absence of centre white lining can increase uncertainty for drivers and removal of the white lining on moderately narrow roads through rural villages has been suggested as a means of reducing vehicle speed. Results for the village of Stiffkey in Norfolk, where speeds were already low, indicated a slight reduction in mean speed when the centre white line was removed (Kennedy & Wheeler, 2001a). Unpublished research indicated a reduction of 7mph in mean speed in Starston, another Norfolk village, when the centre white line was removed. Yagar and Van Aerde (1983) found that the addition of a centre line had little effect on driving speeds. Removal of white lining has been successful in a number of villages in Wiltshire (Debell, 2003).

4.42 An idea tested in Drenthe in Holland (De Waard et al, 1995) combined a novel edge treatment with other measures. It was intended to reduce speed variance on rural roads with a lot of slow-moving farm traffic. The road was effectively narrowed by making it uncomfortable for occupants of cars when driving at over 50 mph unless they kept to the centre of the lane, whilst larger vehicles were not affected. The white edge lines were replaced by 4m long rectangles of rough surface (chippings) interspersed by 4m gaps where the road surface remained unchanged. Rough surfacing was also used between the white dashes in the widened centre line. The mean speed of subjects in an instrumented car was reduced by up to 3kph.

4.43 Marked speed limit roundels on the road surface can be used to complement signing as a reminder of the speed limit. They are usually adopted in conjunction with other measures, for example at a village gateway ( see para 4.60). In the absence of other measures, 30 mph roundels were not found to have any effect, but 40 mph roundels reduced speeds by 3 mph (Barker, 1997, Barker & Helliar-Symons, 1996).

4.44 Coloured road surfacing is commonly used in two ways. The first is to emphasize a traffic calming feature, or to warn of a junction. A series of buff-coloured bands incorporating a SLOW marking at an isolated development on a rural road was found to be effective in reducing mean speeds by 6mph on a driving simulator (Taylor et al, 2002b). The second use of coloured surfacing is to delineate the road space (e.g. by use of cycle or bus lanes).

4.45 Rough road surfaces could include roughness caused by road surface materials (e.g. a brick or cobbled road) or simply "pot-holes" in the road surface. The rougher the road surface, the greater the noise and vibration, and thus driver discomfort, caused. Drivers can be expected to make a rational decision to reduce speed based on utility when exposed to such discomfort when driving. However, care needs to be taken to ensure that road surfaces are not so rough that they result in damage to vehicles or decreased levels of safety due to too much of an adverse effect on the driver. In addition, rough road surfaces can cause problems for cyclists, and increased noise.

4.46 Research evidence shows that rough road surfaces are effective in reducing speeds (e.g. De Waard et al., 1995; Slangen, 1983; Te Velde, 1985; Van de Kerkhof, 1987, Kennedy & Wheeler, 2001a, Wheeler et al, 1997). Slangen (1983) suggested that as much as a 14-23% reduction in mean speed can be obtained due to rough road surfaces, whilst Van de Kerkhof (1987) stated that roughness of a road surface is the most influential factor in determining mean speed. Wheeler et al (1997) found a large reduction in mean speed when imprinted surfacing was combined with prominent visual measures at a gateway. Kennedy & Wheeler (2001a) reported a reduction of about 4mph in mean speed with imprinted surfacing; there was a change in character of the noise generated by vehicles on the imprinted surfacing compared with tarmac.

Rumble devices

4.47 Rumble devices are small raised areas across the carriageway with a vibratory, audible and visual effect. Rumble strips can be laid out in a single group or in a series of groups, usually with decreasing spacing between the groups. They act as alerting devices rather than causing discomfort. As a result, speed reductions at rumble devices tend to be small than at physical measures such as road humps (Webster & Layfield, 1993, Barker, 1997). Rumble devices can be noisy and are therefore unsuitable near to housing. Their effect tends to lessen over time, since there is less discomfort when they are traversed at higher speeds. They are often used in conjunction with other traffic calming measures.

4.48 More recently, a rumble device (Rumblewave) that alerted the driver but did not create noise nuisance for residents has been trialled (Watts et al, 2002, DfT, 2005).

Junction specific measures

4.49 Acceleration and deceleration lanes are used on high-speed roads to facilitate joining or leaving the road at priority junctions. Similarly, it may be appropriate to provide a right turn lane for traffic turning off a high speed road at these junctions.

4.50 Transverse yellow bar markings with reducing spacing in a reverse exponential pattern have been used on the approaches to dual-carriageway roundabouts and were shown to reduce accidents by about 50% relative to control roads (Helliar-Symons, 1981). The aim was to make drivers think they are travelling faster than they really are, and Denton (1973) recorded a reduction of 13kph in mean observed speed. However, Jarvis (1989) concluded that, although the markings do reduce approach speeds, they appear to act as a hazard warning device rather than through manipulation of drivers' visual fields.

4.51 Haynes et al (1993) found a reduction in accidents of 15% when yellow bar markings were tested on motorway off-slips at grade-separated roundabouts, but the result was not statistically significant. It was anticipated that the effect would be less than that found by Helliar-Symons (1981) for at-grade roundabouts since drivers would already have made the conscious decision to leave the motorway.

4.52 Barker (1997) found little change in mean observed speed at 100m from the junction at sites with coloured bars on the minor arms of rural crossroads, whilst Meyer (2001) obtained reductions of up to 4kph on a high speed road in the US with white painted bars of different widths and patterns.

Bend specific measures

4.53 An advance warning sign, vehicle-activated or not, is normally used to indicate a severe bend. Double white lines used with hatching for channelization may be particularly desirable on bends in order to reduce lane width, increase the separation between the two directions of traffic and reduce the risk of head-on collisions. However, the increased segregation does not necessarily reduce speeds (e.g. Kennedy & Wheeler, 2001a).

4.54 Where a bend is severe, chevron signs are used on the bend itself. Research using the TRL Driving Simulator has shown that one large sign is more effective than four smaller ones (Taylor et al, 2002b)

4.55 Skid-resistant surfacing can give a warning of a bend and also enables drivers to travel round it more easily. However, anecdotal evidence suggests that it can lead to drivers continuing to speed, knowing that they are afforded a greater safety margin than where there is no such surfacing.

Shared space and Quiet Lanes

4.56 The concept of shared road space, whereby roads are designed to cater for pedestrians and cyclists as well as motorists, originated in the Netherlands (e.g. the "woonerf") where large reductions in the number of accidents, particularly involving pedestrians and moped riders have been reported (Alink, 1990) following the implementation of such road environments.

4.57 Recently, this concept has been introduced to the UK as Home Zones, where speeds are limited to 15mph and no one type of road user has priority. However, this concept is not really applicable in rural areas, except possibly to minor roads in villages.

4.58 Quiet Lanes also use the concept of shared road space. They are narrow, single track country lanes in a rural area that form a network and are usually subject to the national speed limit (60 mph). They have low speeds and low flows and are intended to encourage shared use by vehicles, cyclists, pedestrians and equestrians. Traffic calming measures are kept to a minimum (and are usually non-existent over much of the network). The idea is to change the hearts and minds of local residents, persuading them to slow down or drive more carefully on the lanes.

4.59 Kennedy & Wheeler (2004a, 2004b) monitored traffic flows and speeds, numbers of non-motorised road users and public attitudes in two pilot areas. They found that there was a small decrease of about 10% in flow relative to control sites, but little change in mean speeds, which were already low because of the narrowness of the lanes and the limited forward visibility.

Villages

4.60 Measures on the approach to a village and/or a gateway are a way of informing drivers of a transition from one type of environment to another (where a different type of driving behaviour might be required). A gateway can involve using a variety of traffic calming measures to reduce vehicle speeds (e.g. signs, emphasis of speed limit sign by use of yellow backing board, countdown signs, roundels, coloured road surfaces, dragon teeth to create a visual impression of narrowing, physical road narrowing etc).

4.61 A number of TRL research studies have investigated the effect of different village gateway schemes on vehicle speeds (e.g. Wheeler, Taylor & Payne, 1993; Wheeler, Taylor & Barker, 1994, Wheeler & Taylor, 1999). Gateways with simple signing and marking measures may reduce mean speeds by about 1-2mph, whilst more comprehensive gateway measures with high visual impact (e.g. coloured road surfacing and dragon teeth) may reduce mean speeds by 5-7mph. When physical measures have been used at gateways (e.g. narrowing), even greater reductions in mean speeds have been found, up to about 10mph.

4.62 It should be pointed out that inhabitants of rural villages often object to the measures with the greatest visual impact as being too intrusive (e.g. red surfacing is often a source of complaint precisely because it is visually intrusive).

4.63 Measures need to be continued beyond the gateway in order to maintain speed reductions through the village itself. These measures are often similar to traffic calming measures used in urban areas, although many physical measures will only be suitable for minor residential roads.

Self-explaining roads

4.64 One reason people often give for driving faster than posted speed limits is that they were not aware of the posted speed limit, and therefore did not know the appropriate speed at which they should have been travelling (Cameron, 1978, 1980; Corbett & Simon, 1992, AA Foundation, 2001). The idea of Self-Explaining Roads (e.g. Kaptein & Classens, 1998; Theeuwes, 1998) is that roads are designed to indicate to drivers the speed at which they should drive through the features of the road itself (alignment, roadside features, etc).

Psychological traffic calming

4.65 Psychological traffic calming involves the use of traffic calming measures other than speed humps and chicanes. A Scottish Executive study (Scottish Executive, 1999) attempted to identify the underlying principles behind natural traffic calming, using ten small or medium towns on through-routes in Scotland as case studies. It was concluded that examples of natural traffic calming tend not to rely on a small number of key features but that drivers are influenced by a large number of different cues. The main components were identified as follows:

• the road corridor as a whole should be considered
• measures should fit the local environment
• location of measures should be matched to natural transitions (e.g. from rural to urban)
• measures should be matched to speed

4.66 TRL has undertaken major studies on behalf of the Highways Agency (2002) and DfT (current). The research for the Highways Agency (2002) suggested various design elements should be considered when developing traffic calming schemes, of which the following are relevant to rural roads:

• Context e.g. roadside type
• Scale e.g. road width and complexity
• Proportion (height of enclosing features such as buildings or trees)
• Horizontal and vertical alignment
• Landscape
• Colour and material of surfacing

4.67 The DfT research (to be published shortly) included a review of psychological principles (Elliott et al, 2003), questionnaire surveys, a simulator trial and on-road trials. The latter showed substantial reductions in mean speed for a psychological scheme in the village of Latton in Wiltshire.

Lay-bys and passing places

4.68 Unpublished research by TRL for the Highways Agency suggests that lay-by accidents, although very rare, are more likely when the lay-by is sited on a bend, or where drivers turn into a lay-by on the opposite site of the road, either to park or as a U-turn.

4.69 In some countries, e.g. the USA, 'turn-outs' are provided at regular intervals. The main purpose of these is to allow slow-moving vehicles a place to pull in and let faster traffic pass. It is a requirement that these vehicles use turn-outs if a queue has built up behind them.

4.70 Research by Walker et al (1964) showed that passing places on single track roads need to be located at frequent intervals to avoid the need for drivers to reverse. These roads are only acceptable where flows are very low, as they have considerably less capacity than two-way roads and can lead to road rage and poor driver behaviour (Kennedy et al, 2004).

Safety barriers

4.71 Safety barriers, which have been used on rural roads for many years, are designed to deflect errant vehicles back onto the road and therefore avoid collisions with roadside hazards such as trees or large sign posts. They are also used extensively in the medians of dual carriageways and motorways. Safety barriers are, therefore, mainly deployed to prevent the run-off-the-road type accidents on single carriageway roads and to prevent head-on collisions on dual carriageways.

4.72 Safety fences operate best when the errant vehicle hits the barrier at an angle of less than 20 degrees and this can be a problem on winding roads. The use of passively safe posts for large signs can reduce the need for safety barriers.

4.73 Research from Scandinavia has shown that the majority of fatal run-off-the-road accidents occur fairly close to the edge of the road (see VTT Finland, 1998). The explanation given is that this reflects the fact that most vehicles roll and/or hit an obstacle quite soon after leaving the road (therefore at higher speed). The most common roadside hazard over much of the network in Scotland will be trees. Whilst these contribute to the attractiveness of the scenery, keeping the verge clear of large trees will reduce the need for safety barriers.

4.74 Safety barriers are not without problems. Motorcyclists in particular are more vulnerable to serious injury and death if they collide with a safety fence which has exposed posts. Work is ongoing to develop safety barriers which can reduce the risks posed to motorcyclists (DfT, 2004f).

4.75 As already mentioned, safety barriers are currently used in the median of all motorways and most dual-carriageways in the UK, to reduce cross-over accidents. However, recent trials in Sweden indicate that there may be potential to apply their use to separate opposing traffic streams on some single carriageway roads (see paragraph 4.79 below).

2+1 layouts

4.76 As noted in Chapter Three, single carriageways are more dangerous than dual carriageways. The reasons for this safety difference are numerous, but two of the key reasons are, the lack of separation between opposing traffic streams, and the potential for overtaking traffic to come into conflict with opposing traffic (at potentially catastrophic speeds). The OECD (1999) recognises that one way to reduce head-on collisions associated with overtaking manoeuvres is by providing conflict-free overtaking opportunities. One way of doing this is the '2+1' layout.

4.77 The '2+1' layout consists of two lanes in one direction and one lane in the opposing direction with the 'extra' overtaking lane ideally alternating in direction every 1-2.5km. The aim is to provide safer, regular overtaking provision on single carriageway roads (and therefore reduce accidents associated with overtaking) which do not justify the considerable extra costs and land-take associated with building traditional dual carriageways. This layout has been used for many years in Germany and has recently been introduced in Scotland on the A68 at Soutra Hill and on the A9 near Newtonmore.

4.78 On hills, climbing lanes (which are similar to 2+1 layouts) can be provided to allow faster vehicles to overtake ones that are slow-moving, thereby reducing delay and driver frustration.

4.79 In Sweden, the authorities have installed over 200km of roads with 2+1 layouts and have recently introduced the addition of cable barriers in the median to physically separate opposing traffic streams (Bergh et al, 2003). The Swedish authorities have developed extensive criteria for the introduction of such layouts which also include the introduction of cable barriers at the roadside to reduce run-off-the-road accidents. The initial results indicate that accident severity is significantly reduced (killed and seriously injured down by 30-50%).

4.80 However, the use of cable barriers as a median is not without problems. Bergh et al (2003) report that maintenance costs are increased, safety during maintenance work is affected and truck breakdowns in the single lane sections can cause lengthy delays.

Enforcement

4.81 The OECD (1999) asserts that " police enforcement is effective in reducing crashes". It highlights a number of conclusions regarding traffic enforcement on rural roads. These include the fact that repeated enforcement (in contrast with 'blitz' campaigns) and random enforcement produce longer 'halo' effects. Also, it recommends that automatic enforcement technologies targeted to address the principal rural road accident types may be the best approach in rural areas.

4.82 Lonero et al (1995) review various research studies on enforcement and suggest that for enforcement measures to be successful at modifying road user behaviour, the following attributes should be sought:

• Sanctions must have 'real bite' and be kept up-to-date with public perceptions of probable apprehension
• The 'theory of games' pattern where the strategy of the enforcement authorities influence the behaviour of the drivers (and vice versa) needs to be broken
• Crash investigation should contain improved behaviour analysis

4.83 Cliff (2003) highlighted the particular problems of road policing in rural areas in New Zealand. In particular, he points out that a small number of officers must cover large areas and carry out enforcement action against drivers who are often known to them personally. The paper reports on an initiative in one district of New Zealand where an integrated approach has been adopted which includes: improved crash reporting, individual officer performance monitoring and intelligence based deployment according to risk (with the emphasis on drink-driving). The conclusion was that the effectiveness of the approach could be demonstrated through improvements in seat-belt use, open road speeds, hospital admissions, reductions in fatal crashes and the proportion of alcohol-related crashes. Overall, the initiative is seen as a best-practice model for rural road policing.

4.84 Winnett (1994) suggests that enforcement measures such as speed cameras can have an important effect on the faster drivers, although it is recognised that their effect tends to be localised and repeated measures may be required. It should be remembered that speed cameras can not provide enforcement against drink-drivers, aggressive or dangerous drivers amongst others. Therefore, some level of traffic police enforcement is still required to back up publicity and engineering measures.

Speed cameras

4.85 Speed cameras are known to reduce speeds and accidents substantially. However, the exact quantification of the benefits is subject to debate since cameras are generally installed at accident blackspots and claims for their success do not always take this into account. In Great Britain, the fines revenue generated by cameras can be ring fenced for road safety but in order to use this revenue, there are strict rules governing camera installation (DfT, 2004c).

4.86 Speed camera housings and/or signs alone have been shown to reduce drivers' travelling speeds, even in the absence of speed cameras themselves (Corbett & Simon, 1999). These measures are highly likely to influence drivers' speed by increasing perceived likelihood of enforcement. However, the deterrent effect of housings and/or signs is less than that of cameras.